You've covered quite a bit in this section and should be gearing up to start conducting your own hypothesis testing! Before moving on to that exciting realm, take a minute to review some of the key takeaways.
Remember that the section began where the last left off, examining the relationship between
- Setting alpha equal to 0.05 (or 0.01)
- Requiring power values of 0.8 or greater
After a thorough investigation of this relationship, you then also saw an alternative t-test, Welch's t-test which can be used for comparing samples of different sizes or different variances. While the formula was a bit complicated, the most important piece to remember is that when the assumptions that sample size and sample variance are equal for the two samples is violated, use Welch's t-test rather than the Student's t-test.
Aside from ensuring that the assumptions of a t-test are met, it's also important to know how type I errors are compounded if you perform multiple tests. This is known as the multiple comparison problem and you saw that type I errors compound under multiple tests. So while the probability of a type I error is equal to
Remember that simply observing a low p-value is not meaningful in and of itself. There are a number of factors to take into consideration when interpreting the results of a statistical test, from alpha, power, sample size, effect size, and the formulation of the problem itself. Good hypothesis testing requires careful thought and design.